Tackling Class Imbalance with Ranking

The dataset comes originally from UCI Machine Learning. The multiclass datasets were transformed in binary classification as mentioned in the paper. Ranking methods were applied to improve class imbalance. The datasets are divided in 30 folds so that other class imbalance methods can be compared to the methods in the paper. The code used in the paper is also provided.

Data ja resurssit

Implementation of the modelsbz2
This package includes Python implemented by the authors of the models used in...
Tutki
- Lisää tietoa
- Lataa
Datasets split by foldsbz2
The dataset comes originally from UCI Machine Learning. The datasets were...
Tutki
- Lisää tietoa
- Lataa
Extended results tablePDF
The extended tables include more metrics for the models evaluated in the paper.
Tutki
- Lisää tietoa
- Lataa

Lisätietoja

Kenttä	Arvo
Lähde	http://archive.ics.uci.edu/ml/
Laatija	Ricardo Cruz; Kelwin Fernandes; Jaime S. Cardoso; Joaquim F. Pinto Costa
Viimeksi päivitetty	maaliskuuta 15, 2019, 11:03 (UTC)
Luotu	helmikuuta 20, 2017, 15:15 (UTC)
DOI	https://doi.org/10.25747/R02P-AP28
dc.Date	July 2016
dc.Format	.pdf; .bz2
dc.Format.Extent	14MB
dc.Language	En
dc.Publisher	INESC TEC
dc.Relation	Cruz, R., Fernandes, K., Cardoso, J. S., & Costa, J. F. P. (2016, July). Tackling class imbalance with ranking. In Neural Networks (IJCNN), 2016 International Joint Conference on (pp. 2182-2187). IEEE. DOI: 10.1109/IJCNN.2016.7727469 ; http://ieeexplore.ieee.org/document/7727469/